Summary
Rodrigo Castellon is a machine learning scientist with nine years of experience building research-driven ML systems, currently working on foundation models for Tesla Autopilot in Palo Alto. He combines a strong academic pedigree from Stanford with hands-on research that produced CVPR and ISMIR publications and a TPDP presentation on differentially private synthetic data. Rodrigo has a track record across industry labs—Bloomberg, ByteDance, Amazon—where he delivered practical advances like the first performant generative model for private tabular synthesis and a patent in audio/music ML. His research spans generative diffusion models for dance-from-music, music-audio representation learning, and realistic MIDI synthesis, showing a rare blend of creativity and engineering rigor. Comfortable moving ideas from papers to production, he focuses on scalable, privacy-aware generative solutions for complex audio and sensor domains. Notably, his work often intersects creative ML (dance and music) with privacy and safety considerations relevant to autonomous systems.
9 years of coding experience
3 years of employment as a software developer
Master of Science - MS, Computer Science (Systems Track), Master of Science - MS, Computer Science (Systems Track) at Stanford University
English, Spanish, Chinese